sophiabits

8 ms·

31.
by sophiabits 3y ago
A case study: The story goes that Snapchat intentionally made their UI hard to use as a way of filtering out adults. An influx of adults would’ve made their app a lot less “cool” in the eyes of their target audience (teens) HN’s UI definite
32.
by sophiabits 3y ago
In C++ there’s an attribute called nodiscard which solves this problem assuming the API you’re consuming is written well. When you apply this attribute to a function, the compiler emits a warning when that function’s return value isn’t used
33.
by sophiabits 3y ago
https://sophiabits.com/blog I started a few years ago but only found a good rhythm in the past ~year. I am 75% of the way to my (arbitrary) goal of hitting 100k written words. Lots of posts on architecture, AWS, and perform
34.
by sophiabits 3y ago
Something really nice about type prefixes in IDs is it makes “Global Object Identification” [1] in GraphQL super straightforward. The node query can simply inspect the ID and immediately know where to fetch the object from, whereas with “re
35.
by sophiabits 3y ago
I have a thin JavaScript package published which generates type-prefixed KSUIDs. KSUID is a great format assuming you can spare the extra bits (which is most people) [1] https://github.com/sophiabits/resource-id
36.
by sophiabits 4y ago
Hmm I could be wrong about this. My personal experience with the Index was that glasses were really uncomfortable, and I've heard from friends that they've had issues with other headsets. Maybe we all have large glasses? :) I didn
37.
by sophiabits 4y ago
Yeah, good question. At a high level I think the best analogy is React: when I first saw it, I wasn't entirely sure about mixing in HTML with my Javascript. It seemed strange. Now that I've been using React for years, however, col
38.
by sophiabits 4y ago
That lines up with my AI experience, too. These days the entirety of my AI experience is sitting on huddles with juniors/intermediates who will type out half a line of code and then wait for Copilot ad infinitum. It can be a bit annoyi
39.
by sophiabits 4y ago
I used Tailwind for the first time about a month and a half ago, reworking my blog from `styled-components` [1]. I had been avoiding it because--as other people have said--it looked really ugly and unmaintainable. I was pleasantly surprised
40.
by sophiabits 4y ago
IMO the hardware simply isn't good enough for the mass market yet. * Motion sickness is still a really big problem * If you're short-sighted and wear glasses, then the picture will be blurry for you. The original PSVR worked aroun
41.
by sophiabits 4y ago
The exact data on how much search traffic converts varies wildly from business to business so it’s hard to give a ballpark figure. The big win from SEO traffic that I’ve seen is that CAC for these customers tends to be a lot lower than for
42.
by sophiabits 4y ago
React & friends are fast to build apps in and are quick to onboard new engineers on. Developer experience has been prioritized over user experience.
43.
by sophiabits 4y ago
Perhaps we need an “LLM Block” browser extension? :)
44.
by sophiabits 4y ago
Grammarly is an OpSec nightmare that’s somehow managed to slip under most people’s radar. I know folks who won’t use the Okta browser extension because of the extensive permissions it asks for, but will happily use Grammarly on everything.
45.
by sophiabits 4y ago
> then, does anybody actually write pure k8s yaml? You can think of plain k8s yaml as a sort of machine code for k8s. You _can_ write yaml by hand, but it’s not generally how things are done outside of initially learning how k8s works. T
46.
by sophiabits 4y ago
At one point in time, typing “H” into Google New Zealand would yield Ian H. Witten (author of Managing Gigabytes) as the first result. His contributions to our field are massive. University of Waikato also work on a digital library software
47.
by sophiabits 4y ago
Exactly! I was thinking of shadowing in Rust when I wrote my original comment. My day job is predominantly in Typescript and a lot of code winds up reading significantly worse than it needs to. A common pattern for me is unique-ifying some
48.
by sophiabits 4y ago
The very last paragraph sums it up pretty nicely imo: > You'll have to be more creative when coming up with variables that don't shadow existing ones (which, for me, generally means using more obscure names). The last adjective
49.
by sophiabits 4y ago
I agree. None of those people you’ve mentioned would recognize the concept of “work-life balance” at all. If you want to be exceptional at something then there’s no substitute for putting more effort in. In fact, if you look at Google books
50.
by sophiabits 4y ago
As it exists today I’d agree with you that GPT is just a tool. I think without some architectural improvements it will likely always be a tool—predicting the next token probably accounts for some part of human intelligence, but I find it di
51.
by sophiabits 4y ago
Hopefully this does end up happening. One of the reasons I personally don’t mind Apple’s 30% IAP take is because going through Apple guarantees I won’t need to deal with any dark patterns—cancelling is one or two clicks in the settings app,
52.
by sophiabits 4y ago
I recently paired with someone who was running into issues configuring some settings in the Facebook developers console. I got them to demo the issue for me, and they /immediately/ closed out of the error dialog when it popped up.
53.
by sophiabits 4y ago
Wave was so far ahead of its time, it’s almost unbelievable
54.
by sophiabits 4y ago
If Javascript is enabled there’s ultimately very little that can be done to prevent fingerprinting. If you don’t want to be fingerprinted then only allowing JS to run on allowlisted websites is the only way to truly be safe
55.
by sophiabits 4y ago
> In Rust there's a focus on ergonomics. How can we make this easier to use? How can we make the tooling nicer? rustc’s error messages are _unbelievably_ nice. It’s shocking how much better they are, even compared to other modern la
56.
by sophiabits 4y ago
At my university the language used for teaching was predominantly Java, with Python or C# used in some papers. There was an optional operating systems paper which had you writing some assembly to run on an in-house RISC system [1], but a sm
57.
by sophiabits 4y ago
They’ve been achieving ~two 9s for the past few months. Given the surge in traffic they must be receiving, that’s honestly pretty decent.
58.
by sophiabits 4y ago
I generally think deploying first and asking questions later is the way to go—when you run in to problems, you get a good opportunity to debug your software development processes to try and prevent similar issues down the line. Something’s
59.
by sophiabits 4y ago
I suspect that humans have an accuracy lower than 99.999%, and are similarly capable of producing confidently incorrect results. GPT has a lot of hype and hysteria around it, but demanding 100% accuracy from it is a bit over the top imo. It
60.
by sophiabits 4y ago
That's a fair point :) I was quite fortunate to have university professors with a decent bit of industry experience, and I /think/ they likely would have been up to the challenge but that's far from the norm.

More ›